Probabilistic Mapping by Fusion of Range-Finders Sensors and Odometry

Abstract

In this work we propose a method for mapping with spatial representation of the environment based on occupancy grid. Our method incorporates a probabilistic model for sonar that consider the uncertainties inherent to this type of sensor as well as for the accumulative errors caused by odometry system of the robot. With this modeling, the quality of the map gets influenced by the uncertainty given by the odometry system indicating the actual trustworthiness of the sensory data collected by the robot system. Once a map gets constructed, the robot can use it to perform other high-level tasks as navigation, path planning, and decision taking, between others.